NVIDIA CEO Jensen Huang recently praised Tesla, stating that its Full Self-Driving (FSD) system is currently the most advanced system. Tesla's FSD also happens to use NVIDIA chips.
In an interview, he mentioned, "Tesla is far ahead in autonomous driving cars. The revolutionary aspect of Tesla's 12th version of fully autonomous driving cars is that it is an end-to-end generative model."
"It learns how to drive end-to-end by watching videos (surround videos) and uses generative artificial intelligence to predict paths, as well as understand and drive cars. Therefore, this technology is truly revolutionary, and the work done by Tesla is incredible," he added.
It is understood that Tesla's latest FSD version 12 is currently in the testing phase, offering a 30-day free trial earlier this year to new car owners. The current charge for FSD is $99 per month or a prepayment of $8,000.
Tesla reported in April that since its debut in March 2021, FSD has accumulated over 1.3 billion miles of driving.
However, it is important to note that FSD is still considered a Level 2 autonomous driving system, meaning it requires supervision and is subject to recalls and government scrutiny of its capabilities.
Nvidia Benefits
According to Nvidia's first-quarter earnings report, its automotive business revenue was $329 million, a small fraction compared to its $22.6 billion data center business revenue, but showing a 17% increase quarter-over-quarter and an 11% increase year-over-year.
Nvidia's CFO Colette Kress stated during the company's earnings call that she expects automotive to become the "largest vertical market within data centers this year" and could potentially be a multi-billion dollar business for the company.
When discussing the future development of autonomous driving technology, Huang Renxun predicted that one day in the future, "every car" will be equipped with some level of autonomous driving capability. This advancement will greatly drive the demand for computing power, with Nvidia's chip technology playing a core role in this aspect.
"This technology is very similar to large language models, but it requires a massive training infrastructure," he said. For instance, to enhance Tesla's ability to process vast amounts of data, Nvidia mentioned that it helped Tesla expand its FSD training AI cluster to 35,000 Nvidia H100 GPUs.
During Tesla's earnings call last month, Musk revealed that by the end of 2024, Tesla will have 85,000 Nvidia H100 GPUs for training artificial intelligence.
Apart from Tesla, Nvidia's other customers in the automotive industry include Mercedes-Benz, Jaguar Land Rover, Volvo, Hyundai, as well as Chinese electric vehicle newcomers like BYD and NIO.
Wall Street Optimism
Not only Nvidia itself, but Wall Street also holds a positive outlook on its automotive business.
In a report released last Thursday, JPMorgan analysts raised Nvidia's target stock price from $850 to $1,150, predicting a 20-30% annual growth rate for the chip manufacturer's data center business.
The report also specifically mentioned an expected "incremental automotive revenue of around $14 billion" for Nvidia in the next three to four years.
